Train the agent slides

Train the agent slides

A browser slide deck on how AI agents are trained with TRL and OpenEnv, from SFT on agent traces to GRPO and RL in environments. Made for developers learning agent training, with PDF and PPTX copies.

Made by sergiopaniego on Hugging Face. These notes are Mindket's, written from its Space page and README; the app's own files and license are on Hugging Face.

About

What it is: This Space is a slide deck about training AI agents, not only models. It covers agent training with TRL and OpenEnv, and every method in the deck matches an episode of the Training Agents live series. It starts with supervised fine-tuning on recorded agent traces, then distillation, GRPO and reinforcement learning inside environments, and ends with training through the agent's wrapper code.

Who it suits: Developers and students who want a guided tour of agent training methods, and speakers who want a deck to read or adapt.

Industry | Who uses it | How they use it | Example
AI engineering | ML engineers | Get an overview of agent training methods | Step through the GRPO and environment slides
Research | Reinforcement learning researchers | See how the methods fit together | Read the slide on training through the agent's wrapper code
Education | Students | Follow a talk at their own pace | Open the deck and press n for speaker notes
Teaching | Speakers and trainers | Reuse the deck as a talk reference | Download the PDF or PPTX copy

If you want code to train an agent right now, a slide deck shows the ideas and does not run training.

Where to use it: It needs only a browser.

• A web browser, on the Space's page at Hugging Face with nothing to install
• A downloaded copy on Windows, macOS or Linux, opened in any browser
• Google Slides, Keynote or PowerPoint, using the PPTX copy

Use case | Who | What they get back
Learn the methods | Developers | A slide per idea, one method per episode
Present the talk | Speakers | Fullscreen slides with speaker notes
Share offline | Teams | A PDF or PPTX copy of the whole deck

How to set it up: Viewing it online needs nothing. A local copy needs two things.

• Python and pip
• Any web browser

1. Install the Hugging Face tools.
2. Download the Space into a folder.
3. Move into that folder.
4. Open index.html in your browser.

pip install -U huggingface_hub
hf download sergiopaniego/train-the-agent-slides --repo-type space --local-dir train-the-agent-slides
cd train-the-agent-slides

How to use it: Open the deck and move through it with the keyboard. Settings offer two buttons, PDF and PPTX, that hand you copies made in advance. The last slide carries a QR code that points back to the deck. The README lists two commands for developers who work on the deck itself.

npm install
npm run dev

Tool | What it does
Right arrow or Space | Goes to the next slide
Left arrow | Goes back one slide
t | Switches between dark and light
f | Turns fullscreen on or off
n | Shows the speaker notes

Example: You open the Space, press f for fullscreen and step through with the right arrow. On a slide you want to study, you press n to read the speaker notes, and press t if the light theme is hard on your eyes.

For newcomers and experts: An agent is a model that takes actions in a loop. Training the agent means improving that whole loop. Start at the first slide and read one method at a time.

The deck groups the methods by what the model learns from. Traces and distillation teach by example, while GRPO and RL in environments teach by reward. Figures sit in sandboxed frames, which is why the PDF and PPTX are built ahead with a real browser pass over every slide, using npm run build and npm run export.

Start from SFT on agent traces
Add distillation
Train with GRPO
Run RL in environments
Train through the wrapper code

Good to know: The PPTX holds one full-slide image per slide, so slides import as pictures, which Google Slides, Keynote and PowerPoint accept.

• License: Apache 2.0
• Cost to run: none named
• Data it reaches: only the deck's own files
• Not for: exporting in the page on demand, since PDF and PPTX copies are prebuilt

On Hugging Face, 13 November 2026

• Likes: 20
• Built with: static
• License: apache-2.0

Links

• Space: https://huggingface.co/spaces/sergiopaniego/train-the-agent-slides
• Files: https://huggingface.co/spaces/sergiopaniego/train-the-agent-slides/tree/main

Try it

Open the Space in your browser, or copy it to your own Hugging Face account from its page. Nothing in this kit runs by itself.